Telomeres

Human telomeres consist of long tandem arrays of the hexanucleotide TTAGGG. Moyzis and colleagues isolated this sequence in 1988 and showed by in situ hybridisation that its major clusters sit at the telomeres of all human chromosomes; the sequence is conserved across many higher eukaryotes [1].

The repeats are bound by shelterin, a complex of six telomere-specific proteins. Three subunits (TRF1, TRF2 and POT1) recognise TTAGGG repeats directly, and three more (TIN2, TPP1 and Rap1) link them together. Shelterin lets the cell distinguish natural chromosome ends from sites of DNA damage; without it, chromosome ends are exposed to DNA damage surveillance and are processed inappropriately by DNA repair pathways [2].

Part of this protection comes from DNA architecture. Electron microscopy of telomeric DNA from human and mouse cells shows large duplex loops, called t loops, formed when the single-stranded 3' overhang at the chromosome end invades the double-stranded repeat array. TRF2 can remodel linear telomeric DNA into t loops, and inhibiting TRF2 causes immediate loss of end protection, with loss of the 3' overhang, activation of the p53 damage response and end-to-end fusion of chromosomes [11].

The DNA-copying machinery of the cell cannot fully replicate the extreme end of a linear DNA molecule, so a small amount of sequence at each chromosome end is lost in each round of copying. Alexey Olovnikov set out this idea in his 1973 "theory of marginotomy", which proposed that incomplete copying of the template margin would progressively shorten chromosomes and could explain the limited division capacity of normal cells [12].

Direct evidence followed in 1990, in a study by Calvin Harley, Bruce Futcher and Carol Greider: the amount and length of telomeric DNA in human fibroblasts fell with serial passage in culture, and possibly also with age in the body [3]. At the time the authors noted that it was not known whether this loss caused cellular senescence [3].

In 1961 Leonard Hayflick and Paul Moorhead reported that normal human diploid cells grown in culture undergo a finite number of divisions and then stop [4, 5]. The limit became known as the Hayflick limit, and the permanent non-dividing state that follows is called replicative senescence.

A causal link between telomere shortening and this limit was established in 1998. Retinal pigment epithelial cells and foreskin fibroblasts, both normally lacking telomerase, were given the gene for its catalytic subunit. Control cells showed telomere shortening and senescence, while telomerase-expressing cells had longer telomeres, kept dividing, showed less staining for a senescence marker and exceeded their normal lifespan by at least 20 population doublings with a normal set of chromosomes [5].

Senescence stops damaged or dysfunctional cells from dividing and is an important brake on the progression of tumour cells [13]. Senescent cells also accumulate in many tissues with age and are thought to disrupt tissue structure and function through the molecules they secrete, a pattern called the senescence-associated secretory phenotype [13]. In a 2011 study in prematurely aging mice, removing cells that expressed the senescence marker p16Ink4a delayed age-related disorders of fat tissue, skeletal muscle and the eye, and removal late in life slowed disorders already under way [13]. This link between senescence and chronic inflammation is one reason telomere biology overlaps with other theories of aging.

Telomerase is an enzyme that adds telomeric repeats to chromosome ends. Carol Greider and Elizabeth Blackburn identified its activity in 1985 in extracts of the single-celled organism Tetrahymena, where it added TTGGGG repeats to synthetic telomere primers independently of the usual DNA polymerase; activity was higher during a developmental stage in which many new telomeres form [14].

In humans, telomerase activity is largely confined to immortal cells and germline tissue. A 1994 survey using a sensitive assay found telomerase in 98 of 100 immortal cell populations and none of 22 mortal ones, and in 90 of 101 tumour biopsies from 12 tumour types but none of 50 normal somatic tissues. Normal ovaries and testes tested positive, while benign tumours such as fibroids tested negative [6]. The authors concluded that telomerase is tightly repressed in normal body tissues and reactivated in cancer, where cells must keep dividing to sustain tumour growth [6].

This dual role shapes research on telomerase as a target. Shay and Wright's 2019 review of three decades of the field describes telomerase inhibition as a potential cancer therapy and lists genetic disorders that cause premature telomere shortening as a major area of discovery [15]. Because telomerase activity is a near-universal feature of tumours [6], products intended to activate telomerase throughout the body carry a theoretical cancer concern.

Telomere length in white blood cells (leukocytes) is the measure most often used in population studies, because blood samples are easy to collect. A 2018 meta-analysis (a statistical pooling of results from several studies) combined 21 published studies with 4 unpublished groups from the Swedish Twin Registry, for 25 studies covering 121,749 people and 21,763 deaths [9].

Across all studies, each one standard deviation shorter telomere length was associated with a hazard ratio of 1.09 (95% confidence interval 1.06 to 1.13). A hazard ratio compares the rate of death in one group with another, so 1.09 means a 9% higher rate; the 95% confidence interval is the range in which the true value probably lies. People in the shortest quarter of telomere length had a 26% higher death rate than those in the longest quarter (95% CI 15% to 38%) [9]. In the Swedish twin data alone (12,083 people, 2,517 deaths), the corresponding figures were 13% per standard deviation (95% CI 7% to 19%) and 44% for the shortest versus longest quarter (95% CI 27% to 63%) [9].

Results differed between studies. The association was weaker in people over 80, and differences in measurement technique and age at measurement contributed to the variation between studies [9]. These are observational associations in groups of people; they do not establish that short telomeres cause death, and a single length measurement gives limited information about one person's risk.

The best-known study linking stress to telomere length examined 58 healthy premenopausal women: 39 mothers caring for a chronically ill child and 19 mothers of a healthy child [10]. Both perceived stress and the number of years spent caregiving were associated with higher oxidative stress, lower telomerase activity and shorter telomeres in blood immune cells. Women with the highest perceived stress had telomeres shorter, on average, by the equivalent of at least one decade of additional aging compared with women reporting low stress [10].

The study was small and cross-sectional, measuring stress and telomeres at the same point in time, so it cannot show that stress caused the shortening. The accompanying rise in oxidative stress and fall in telomerase activity are proposed pathways through which chronic stress could speed telomere loss [10].

In 2013 López-Otín and colleagues proposed nine hallmarks of aging, common features of aging across organisms: genomic instability, telomere attrition, epigenetic alterations, loss of proteostasis, deregulated nutrient sensing, mitochondrial dysfunction, cellular senescence, stem cell exhaustion and altered intercellular communication [7]. The 2023 update expanded the list to twelve, adding disabled macroautophagy, chronic inflammation and dysbiosis. A process qualifies as a hallmark if it appears with age, if accentuating it experimentally accelerates aging, and if treating it can slow, stop or reverse aging [8].

The authors describe the hallmarks as interconnected rather than independent [8]. Telomere attrition feeds directly into cellular senescence through the division limit, and senescent cells in turn contribute to the chronic inflammation described in inflammation and aging [13]. In this framework telomere shortening is one of several interacting processes, not a single master clock of aging.